|Information Architecture for the Small Site - part 2|
Label the content piles
| 10:03 pm on Apr 23, 2004 (gmt 0)|
This post continues on from Part One [webmasterworld.com], where we had separated our website content into several piles of index cards, and created no more than seven piles that were destined for the main menu.
Now comes a second critical point in Information Architecture - choosing the lables for those piles. These will be what the user sees when they look at the website's main menu, and those labels must communicate cleanly and clearly what the user will discover if they make any particular click.
Words are funny things - we think we all mean the same thing when we use the same word, but that's not true -- and it's especially untrue when you go international.
DELIVERY or SHIPPING
For example, working with a UK website to help increase their US business, I found that they were using the word "Delivery", but in the US it would be called "Shipping". When a third party is involved, that's not delivery anymore in the US. Pizzas get delivered, but web orders get shipped.
To a US customer, "Delivery" sounds like you're going to get a bunch of the boys together and drive the company truck right over to the customer's front door. If the site is obviously a UK business, that might create the feeling in a US visitor that you don't do international orders.
That's just one small, slightly "off", note. But if those slightly "off" notes build up, then you've got a confused visitor who will probably leave. And the wrong notes don't need to be US/UK problems, they can just be labels that are not well understood by some percentage of people -- regional differences, levels of education, or just individual quirkiness.
A NEW KIND OF CARD SORT
So here's an approach to choosing your menu labels that can avoid some of the problems ahead of time.
You've already got your content written down on index cards, and all separated into piles in Part One. Now take another group of clean cards and write your proposed labels on them, one menu label per card. Then find someone will to help you out. Shuffle up your content cards, and ask your helper to put each content card in a pile by it's appropriate label card. Listen for their comments and watch for any hesitation -- and don't give hints, just take notes!
It's best to use a couple of helpers at different times, but even one run-through will usually give you valuable feedback. Most likely, you'll end up with some label changes. So test those new labels the same way and keep those card sorts going until you have a set of labels that work -- that create unambiguous names for the piles of content you created.
You may end up moving some content to a different section. You may end up needing to create a completely new separation scheme for the piling up your content. But in the end, you should have some very clear and highly usable labels for your main menu.
There are different styles used in creating menu labels, and there is no one right way. But one of the important issues to resolve is whether you will use phrases or single words. In most cases, you want labels that are roughly similar in length - it makes the menu easier for the eye to scan.
Similarly, try to use the same parts of speech for labels. After all, these are supposed to be sections of roughly equal importance -- so labeling some with nouns and some with verbs and yet others with full phrases often looks poorly structured, or even coherent.
Look at your label choices as a whole unit - your visitors certainly will. Do they accurately represent what the website is all about?
TELLING A STORY
For one site I created recently, I created menu labels that took the visitor through a progression of approach to the business involved -- it was as if the menu told a story, literally.
I used longer phrases than I do most of the time, and one quick read-through told the visitor how to learn the whole story of the business, from why they existed to what they could do for you and how to get involved. That particular site is thriving right now, so it looks like it was a good choice in that case. In another case, it might just look forced, but this was offering professional instruction in a cutting edge discipline. So the visitor needed some motiveation, then education, and then finally conversion.
I wasn't sure about that "story telling" approach to a menu, because as we all know a website is not really a linear progression like a print piece. So what might be the effect of creating a very linear overlay on the website's content?
Well, in this case, with clear menu labels and multiple location cues on every page, the visitor who first arrives in the middle of things could quickly get their bearings -- and the stats for the site show that this is true.
CALLS TO ACTION
Menu labels can also include "calls to action." I've learned that main menu labels like "Register Here", "Learn About XYZ", or "Subscribe to our Newsletter" can be extremely effective in getting the all important conversion -- so much so that I often make sure to save one of my seven spots for an important call to action. Yes, that only leaves me with six categories for all the rest of the content, but the results are where it's at.
Choosing menu labels with care is equally as important as how you slice and dice the content in the first place. Do it well and you create a website that looks easy to deal with, easy to "wrap your brain around" -- even if there are actually great complexities being communicated.
| 2:50 am on Apr 27, 2004 (gmt 0)|
A client just brought me an example of awful menu labels. It's a not-for-profit that has a lot in print about them -- the (fascinating) story of their founding being one example.
But all of this nifty content is hidden in a link called "Literature". Now would you think of clicking on Literature to discover the story of how they got started?
| 4:43 am on Apr 27, 2004 (gmt 0)|
I was one of the ones in Orlando who asked you to commit your talk to "paper." I've gotten far more than I expected. This is a graduate-level course.
| 2:10 pm on Apr 27, 2004 (gmt 0)|
These threads have been a great read Tedster, thanks.
I'm having some difficulty trying to figure out how to apply the rule of 7 to my site.
I'm wondering if the size of the site could be the problem, since you address these threads as applying to small sites.
Any chance you could give an idea of what you mean by "small site"?
I imagine that if this is important for small site, it's equally, if not more important for larger sites.
| 2:15 pm on Apr 27, 2004 (gmt 0)|
These are awesome posts tedster. Thanks for sharing these with us.
And in the spirit of sharing here are two sites which cover some of the same information, albeit not as well as tedster:
Information Architecture for Everyone [evolt.org]
Card sorting: a definitive guide [boxesandarrows.com]
Mods, if these links aren't appropriate, please delete.
| 6:21 pm on Apr 27, 2004 (gmt 0)|
I've successfully applied the "Rule of 7" to sites up to about 1,200 pages. Then it often can get dicey, especially if the main purpose of the site is showing signs of "scope creep". If the original mission of the site is still essentially the same, then the original seven categories still holds up.
But many times, the entire website needs a re-think and a re-design if it has grown to 1,000 or 1,500 pages. If the original size was PLANNED to be that large, then it's usually a better situation.
Still the principle of seven remains important. Are there really more than seven EQUALLY IMPORTANT tasks at the top level of your information hierarchy? Even if it looks like there are, can't some be grouped?
One site I work with began with a plan for about 5,000 pages over 8 areas -- and that initial architecture took over a year to work out with all the stakeholders. The menu became 7 single words distributed around a circle, with an eighth (and most essential) item dead center in the circle. In that way we kept the "equal" choices down to seven, at it really worked nicely.
The center menu choice got the most clicks, but the rest of the menu receives a pretty even distribution. And of the visits that began on the home page, we see only 8% "one hit wonders".
It is a challenge, and I can't pretend to be at the level of a really well trained information architect -- they are pretty rare, but this project had one who taught me a lot about the discipline.
I'm a rare bird, apparently, when it comes to creating web sites. I always am focused on how information is packaged, from the level of individual sentences up to the top-level menu choices.
Others focus on the visual design (in fact, most of my clients want to focus there) or the server sider functionality. I tend to see those items as essential, but supportive rather than primary.
One more thought about the Rule of 7 -- if I break the "rule" I try to do it in a way that allows me to back out gracefully and I measure the results for a couple weeks before committing the future of the site in this hereticla direction. Sometimes under the pressure of time, I just can't see a more ideal solution and I can't wait for the light bulb to flash over my head. But as always on the web, I test and measure.
| 9:16 pm on Apr 27, 2004 (gmt 0)|
Thanks for the additional info Tedster. That does help me get a better grasp. I'm right in that 1,200 to 1,500 page spot at the moment.
The planned vs unplanned, or maybe it's better to say the anticipated vs unanticipated growth, is an interesting aspect of applying these concepts to an established site.
I thought I'd thought this all through before I put up the site to begin with. Sigh... apparently not as well as I might have hoped. Still I did leave a few options open before I started, just for unforseen dvelopments, it looks like that might work in my favor.
Anticipating the unforeseen is something well worth doing at the initial planning stages of a site.
| 9:29 pm on Apr 27, 2004 (gmt 0)|
Anticipating the unforeseen -- one of the big reasons that I love to hit that "sweet spot" of 5 top level choices when a site first launches.
One site I work with offers a series of seminars throughout the year. We really thought hard about whether that deserved a top level menu pot -- and we decided against it. In fact, there is no "index page" for the upcoming seminars at all. They exist in limbo, not really on the site map anywhere.
But they're hardly orphans. With creative linking from the calendar (in a "utility" area) and other relevant areas, we've managed to sell out most of the seminars so well and so early that the company stopped doing direct mail promotion and now relies solely on the web. And the seminar pages get really good search engine rank as well.
Solid Information Architecture requires alignment with the business plan, but some out of the box thinking can keep the site highly usable and effective.
| 12:59 am on Apr 28, 2004 (gmt 0)|
tedster, m'dear.... have you ever considered presenting a college-level course on this sort of thing? I've had profs with PHDs who didn't exposit as clearly as you do!
Really.... how about "WebmasterWorld's Online College of Web-Knowledge". Put up a section for "lecture-fora" with a bit of payment for access to reply function - anyone can read, but if you want to post a question etc, it'll cost a little bit. That would stop people who don't REALLY want to learn what's in the "course", but just want to kibitz - like five bucks per "knowledge-college" forum gives you access to that ONE forum.... Use paypal, easy, accessible....
I'd have paid you or whoever five bucks ANY NUMBER OF TIMES for the stuff you've posted on this so far....
| 11:42 pm on Apr 28, 2004 (gmt 0)|
I'm new to this forum and I joined because I saw a reference to the infamous "No more that 7 rule" in an e-newsletter.
I am not claiming to be an authority on this subject, however, from what I understand not everyone agrees that the 7 +/- 2 rule applies to main navigation bars.
Though Tedster obviously has experience to back up his views, Christine Wodtke for one argues against this kind of application of the "No more than 7 rule" in her book "Information architecture, blueprints for the web".
My understanding is that the 7+/-2 rule was a misinterpretation of George Miller's findings with regards to the short-term human memory. He argued that people can generally only remember 7+/-2 (i.e. 5 to 9) digits or items. However, and this is where the misinterpretation was, this does not apply to the user's ability to select from a given list.
When presented with a list of options, the user simply scans until they find an option that appeals to their interest or needs at that time. If a list (like that of a navigation bar) is always in view, then there is no need to memorise it, so the 7+/-2 rule does not apply.
Perhaps Tedster was referring to some other view or body of research?
| 12:01 am on Apr 29, 2004 (gmt 0)|
|this does not apply to the user's ability to select from a given list |
But it does apply to the user's ability to quickly learn and remember the architecture of the entire site. One of often overlooked parts of IA is giving the user a sense of placement.
|If a list... is always in view, then there is no need to memorise it |
I want my users to *know* where they are and the main sections they can go at anytime. I'd rather they didn't have to scroll through a list of 20 items, to figure out what to do next.
| 1:46 am on Apr 29, 2004 (gmt 0)|
Oh - welcome, Lorry108!
| 12:03 pm on Apr 29, 2004 (gmt 0)|
How might I condense this menu?
My menu, in no particular order,
1) A link for the home page
2) A link for the forum
3) A link used for a calender of events
4) A link used for a photo gallery
5) A search site link
6) A FAQ link
7) A link allowing people to contact me
8) A link used to download zipped information
9) A link used for the actual content of the sight
10) A link that allows visitors to view information submitted by other visitors
I've been mulling over this for about a hour now, ever since I read this post. I see a need for each of these links. I would not want to force a person to search through my FAQ in order to be able to contact me. On the same note, I wouldn't want to force people to dig through content in order to find the downloads their looking for. If a person stops by my website to look up a date on my calender I want them to be only 1-click away.
I like the theory of keep-it-simple-stupid and keeping a slim 7 links does seem more professional. I'm just not sure where to cut the fat.
| 2:12 pm on Apr 29, 2004 (gmt 0)|
I think the point is that you may have to think totally differently. For example "A link that allows visitors to view information submitted by other visitors" may not be a category that works for visitors. Instead you may find that your primary sort would be topics rather than how the content is pulled together.
| 2:28 pm on Apr 29, 2004 (gmt 0)|
It's very difficult, if not impossible to disect your list without knowing the content and the purpose of your site. I think you need to start from part one of this tutorial and review the information contained there.
I never really have a problem with splitting my menu into categories but maybe this stems from the fact that my way into the web was through formal database design which involves similar principles in terms of categorizing information and data.
What does interest me about this particular thread is that of labeling menu items for best effect, I am presently using questions as labels for my menu items on one site.
For example, About Us becomes 'Who are x company?', Widgets becomes 'What widgets are available' and so on until you get to 'how do I buy widgets'. I wonder if anyone else has used this type of approach and has any comments to make.
| 6:58 pm on Apr 29, 2004 (gmt 0)|
Hey thanks, I had read the first post in the other thread but not all of them. I see someone has asked my question already. I think Tedster had a great answer, i'll take my faq, search and contact links and stick them in another menu below my main menu. A perfect 7 item main menu is left!
| 7:29 pm on Apr 29, 2004 (gmt 0)|
|using questions as labels for my menu |
Yes, I've used that approach and it seems to work pretty well, especially on sites that really are about something new and unfamiliar. Direct marketers well know the power of a question to engage the reader.
Some of my study in this area has pointed to using active verbs as menu labels. In that way, each menu item becomes a call to action. I'm not convinced - when I mock up menus that way, they often look contrived. I think the active verb approach is better for inline links, and less so for menu labels.
I'd say that whether using questions, or active verbs, you always need to use common sense as well. If it looks forced and "cute", it probably is, and it won't work.
Most of all, you want the menu to mirror the visitor's mental model of information they want or tasks they'd like to perform. If it looks contrived, it's almost an insult.
| 8:09 pm on Apr 29, 2004 (gmt 0)|
It definitely took some thought and a fair bit of testing to get the wording of the questions right. I ended up with 5 menu items, 4 questions and a final call to action using an active verb.
Will post back when the site has been running long enough and I have more feedback on performance.
The other point is that I don't have a specific 'home page' link. The home page is an integral part of the website, not just an entry page as so many seem to be. My own personal view is that you only need a link back to a traditional home page from every page in the site if the navigation structure doesn't work properly, so that people have a reference point to work from if or rather when they get lost.
| 5:06 pm on Oct 17, 2004 (gmt 0)|
This business of making and labeling your website's "piles of content" is one of the most essential elements you can grapple with - sometimes it's a real bear!
An excellent new discussion about the practical aspects of Information Architecture can now be found here:
Putting Information Architecture into Practice [webmasterworld.com]